DNA evidence alone wasn't a smoking gun in the case against Bruce Ivins as the perpetrator of the 2001 anthrax attacks, say microbiologists and other experts who have read details of the investigation released last week.
Genetic sleuthing was useful in narrowing the list of suspects, they say, but it wasn't conclusive since DNA from bacteria doesn't often carry a unique genetic fingerprint the way human DNA does.
At first, prosecutors seemed to suggest that forensic DNA had solved the case. U.S. Attorney Jeffrey Taylor said science had enabled the government to link the anthrax spores in the 2001 attack to a flask "created and solely maintained by Dr. Ivins" in his federal lab.
But at least eight other anthrax samples gathered from researchers in the investigation carried the same genetic signature as Ivins' batch at Fort Detrick, Md., court documents say.
Investigators used other lines of evidence, including his e-mails, to bolster their case against Ivins, who committed suicide July 29.
Randy Murch, a former FBI expert on bioforensics, said investigators hadn't relied on DNA alone. "Rarely does a case get solved by only forensic evidence," he said. "Here the science is highly informative, but it's also limited."
Some scientists now want more details before they can feel satisfied that the FBI identified the right person.
Dean Boyd, a national-security spokesman for the Department of Justice, said, "Obviously, we anticipated that we would have had to defend the reliability of our scientific findings at trial, and we're confident we would have been able to do so."
When the attacks killed five people just days after 9/11, hundreds of researchers in the United States and abroad were working with anthrax, making them potential suspects.
But the techniques used to connect DNA from crime scenes to human subjects wouldn't work in bacteria.
"Anthrax spends the majority of time as a spore. That's why we don't see diversity in the genome that we see in [other] organisms," said Ted Hadfield, a biologist at the Midwest Research Institute in Kansas City, Mo., and former head of microbiology for the Armed Forces Institute of Pathology.
As spores, anthrax bacteria can exist for years in suspended animation without undergoing the cell divisions that allow mutations to crop up.
With human DNA, investigators can connect an individual to a crime sample by examining repetitive areas called STRs, or short tandem repeats.
Experts liken these to little stutters in the code. Counting the number of times these pieces repeat distinguishes one person's DNA from another's.
Anthrax DNA has far fewer such stutters, scientists say, making it much harder to distinguish one sample from the next.
Until recently, it was hard even to divide anthrax into distinct strains.
In 2000, microbiologist Paul Keim of Northern Arizona University published a landmark paper on repeating regions called VNTRs, or variable number tandem repeats, that occur in anthrax.
Keim, one of the country's top genetics experts, declined to be interviewed, but in several scientific papers he described how he and colleagues developed a system of distinguishing anthrax strains using these VNTRs.
That was one tool the investigators had in 2001. It took them less than a month after the attacks to identify the spores as belonging to an anthrax strain called Ames.
But that still left them with an overwhelming number of potential suspects, since the Ames strain was stored in labs around the world.
The name comes from Ames, Iowa, where veterinarians at the National Animal Disease Center isolated it from sick cattle in 1981. Because this strain was seen as particularly deadly, it became the most common one for laboratory studies, and was used to test the anthrax vaccine given to military personnel.
In the 1980s, a Virginia company called American Type Culture Collection kept samples of Ames anthrax and sent them to labs around the world - including ones in Iraq, which the United States was helping at the time.
Faced with an enormous haystack and an elusive needle, the FBI investigators requested that every U.S. researcher with access to anthrax send in samples of all possible strains. They ended up with 1,000 samples, according to FBI documents released last week.
The one hope they had for solving the case came from the fact that anthrax DNA can occasionally develop single spelling errors called SNPs.
These errors occur in only about one in a million bacteria as they're grown in culture, microbiologist Hadfield said. That could create tiny but recognizable differences among samples.
Finding those spelling errors isn't easy, researchers say. Such a feat would likely require a complete reading of the genetic codes of various samples and comparing them.
In 2001, biologists had just completed the Human Genome Project, for which they sequenced a small sample of human DNA.
At that time, sequencing an organism's entire genetic code would have taken $1 million and months, Hadfield said. But this technology has advanced rapidly, so now it takes just days and about $25,000.
"As we got better at sequencing, we got a better feel for what the SNPs were and where the occurred," Hadfield said.
Investigators eventually found four such SNP-type mutations that distinguished bacteria used in the attacks from samples of the original Ames strain.
Those four mutations were found in only eight of the 1,000 samples under investigation. This subgroup was labeled RMR-1029.
According to the FBI, all the people with positive samples said they had obtained them from Ivins.
But the sample Ivins initially provided in 2002 tested negative for the four key mutations.
In 2004, the documents say, investigators entered Ivins' lab and seized samples, including the "parent" flask that had allegedly supplied the other positive RMR-1029 samples.
The bacteria in that flask allegedly carried the four telltale mutations.
That analysis alone, however, doesn't rule out researchers who worked with the eight samples.
The documents cited other evidence against Ivins, including the source of envelopes used in the mailings and his inability to explain the unusually long hours he spent in his lab before the fall of 2001.
Investigators also presented e-mail messages, allegedly sent by Ivins to colleagues, revealing his struggle with depression and other emotional problems.
Many scientists want more information than the FBI has provided so far - on how it is so sure that Ivins was the right suspect and worked alone, how he carried off the deadly attacks, and why.
Richard Spertzel, a bioweapons expert who worked at the same army lab as Ivins, said the perpetrator had used a sophisticated process to turn the spores into the deadly powder used in the attacks.
He said the machine in Ivins' lab known as a lyophilizer is a common piece of equipment used to dry spores and would not by itself allow someone to create the 1.5- to 3-micron particles used in the attacks.
"He must have used some other new technique that we don't know about," he said.
Spertzel said he found it unlikely that someone acting alone could have created the anthrax used in 2001. "I'd like to see the details behind the hype."